Effect of Abrasive Machining on the Electrical Properties Cu86mn12ni2 Alloy Shunts

materials

Article Effect of Abrasive Machining on the Electrical Properties Cu86Mn12Ni2 Alloy Shunts

Siti Nabilah Misti 1,*, Martin Birkett 1 ID , Roger Penlington 1 ID and David Bell 2

1 Department of Mechanical and Construction Engineering, Northumbria University, Newcastle upon Tyne NE1 8ST, UK; [email protected] (M.B.); [email protected] (R.P.) 2 Vice Chancellors Ofﬁce, Teesside University, Middlesbrough TS1 3BA, UK; [email protected] * Correspondence: [email protected]; Tel.: +44-191-2273763

Received: 10 May 2017; Accepted: 21 July 2017; Published: 29 July 2017

Abstract: This paper studies the effect of abrasive trimming on the electrical properties of Cu86Mn12Ni2 Manganin alloy shunt resistors. A precision abrasive trimming system for fine tuning the resistance tolerance of high current Manganin shunt resistors is proposed. The system is shown to be capable of reducing the resistance tolerance of 100 µΩ shunts from their standard value of ±5% to <±1% by removing controlled amounts of Manganin material using a square cut trim geometry. The temperature coefficient of resistance (TCR), high current, and high temperature performance of the trimmed shunts was compared to that of untrimmed parts to determine if trimming had any detrimental effect on these key electrical performance parameters of the device. It was shown that the TCR value was reduced following trimming with typical results of +106 ppm/◦C and +93 ppm/◦C for untrimmed and trimmed parts respectively. When subjected to a high current of 200 A the trimmed parts showed a slight increase in temperature rise to 203 ◦C, as compared to 194 ◦C for the untrimmed parts, but both had significant temporary increases in resistance of up to 1.3 µΩ. The results for resistance change following high temperature storage at 200 ◦C for 168 h were also significant for both untrimmed and trimmed parts with shifts of 1.85% and 2.29% respectively and these results were related to surface oxidation of the Manganin alloy which was accelerated for the freshly exposed surfaces of the trimmed part.

Keywords: abrasive machining; Manganin alloy; shunt resistor

1. Introduction Shunt resistors have been used as ammeters to measure the flow of electrical current for several decades and offer the advantages of lower cost, low power loss, high stability, and precision of electric resistance across a wide temperature range. This is significant when compared to other current sensing methods such as current transformers, Hall sensors, and Rogowski coils [1]. Shunt resistors are typically manufactured from a Manganin (Cu86Mn12Ni2 wt %) alloy element which is an electron beam welded to two low resistivity copper terminations to permit accurate electrical measurement. Manganin is a commercially available alloy of resistivity 48.2 × 10−8 Ωm [2] and low Temperature Coefficient of Resistance (TCR) of ±15 ppm/◦C[3], offering low and stable resistance across a wide temperature range. It also possesses an extremely low thermal electromotive force (EMF) of 0.1 µV/K at 20 ◦C and excellent long-term stability of electrical resistance. One recent important use of Manganin shunts is in single phase smart energy meters to measure the flow of electrical current. Due to its relatively low cost, the shunt resistor is the preferred current sensing method in this application where it must maintain a stable and repeatable resistance value over a wide current range of 0 to 100 A, across an operating temperature of −25 to +55 ◦C and relative

Materials 2017, 10, 876; doi:10.3390/ma10080876 www.mdpi.com/journal/materials Materials 2017, 10, 876 2 of 10

Materialsair humidity 2017, 10 of, 876 30 to 100% [4]. In order to minimize overall power consumption of the smart energy2 of 10 meter, the room temperature resistance value of the shunt must be as low as possible and typically in the range 100 μΩ to 10 mΩ [5]. Although this low resistance requirement reduces energy usage, it in the range 100 µΩ to 10 mΩ [5]. Although this low resistance requirement reduces energy usage, also causes two significant issues. Firstly, the voltage drop that must be sensed across the shunt is it also causes two significant issues. Firstly, the voltage drop that must be sensed across the shunt substantially reduced, this in turn leads to a requirement to incorporate more accurate measuring is substantially reduced, this in turn leads to a requirement to incorporate more accurate measuring equipment into the energy meter. The second issue, which is the focus of this current work, is that it equipment into the energy meter. The second issue, which is the focus of this current work, is that it is is inherently difficult to manufacture shunt resistors in this low resistance range to the required inherently difficult to manufacture shunt resistors in this low resistance range to the required precision precision and at a reasonable cost. Typical resistance accuracy of commercially available shunts and at a reasonable cost. Typical resistance accuracy of commercially available shunts suitable in this suitable in this application is 100 μΩ to a tolerance of ±5% [6]. When used to measure current flow in application is 100 µΩ to a tolerance of ±5% [6]. When used to measure current flow in the smart the smart energy meter, this tolerance can result in power usage being either over or under calculated energy meter, this tolerance can result in power usage being either over or under calculated by up to by up to 5%. The current method used to reduce this inaccuracy is to calibrate the overall performance 5%. The current method used to reduce this inaccuracy is to calibrate the overall performance of the of the assembled meter. This is not always accurate across the full operating conditions and involves assembled meter. This is not always accurate across the full operating conditions and involves the the addition of compensation software which increases the overall cost of the meter. A more efficient addition of compensation software which increases the overall cost of the meter. A more efficient way way to improve the accuracy of the smart energy meter would be to reduce the resistance tolerance to improve the accuracy of the smart energy meter would be to reduce the resistance tolerance of the of the Manganin shunt resistor itself to less than ±5%. Manganin shunt resistor itself to less than ±5%. There are a number of well-established methods of improving the resistance accuracy of thin There are a number of well-established methods of improving the resistance accuracy of thin and and thick film discrete resistors by removing resistive material and adjusting the geometry of the thick film discrete resistors by removing resistive material and adjusting the geometry of the element element to increase its resistance. The most popular of these techniques are laser trimming, abrasive to increase its resistance. The most popular of these techniques are laser trimming, abrasive trimming trimming with a wheel, or abrasive particles and machining [7]. In the majority of processes, the with a wheel, or abrasive particles and machining [7]. In the majority of processes, the material is material is removed in single or multiple lines cut perpendicular to the flow of current to give high removed in single or multiple lines cut perpendicular to the flow of current to give high rates of rates of resistance change or in parallel with the current flow to give slower rates of change [8]. resistance change or in parallel with the current flow to give slower rates of change [8]. Although these Although these methods have yielded excellent results when adjusting the resistance tolerance of methods have yielded excellent results when adjusting the resistance tolerance of thin and thick film thin and thick film resistors, there has been limited application in the area of bulk metal alloy shunts. resistors, there has been limited application in the area of bulk metal alloy shunts. Recent work by the authors has highlighted the potential of a new trimming approach using Recent work by the authors has highlighted the potential of a new trimming approach using different geometry cuts to adjust the resistance tolerance of shunt resistors [9]. This current study will different geometry cuts to adjust the resistance tolerance of shunt resistors [9]. This current study will investigate the effect of abrasive machining using a square cut geometry on the principal electrical investigate the effect of abrasive machining using a square cut geometry on the principal electrical properties of 100 μΩ Manganin alloy shunts. properties of 100 µΩ Manganin alloy shunts.

2. Experimen Experimentaltal Procedures Procedures

2.1. Materials Materials All samples used in this investigationinvestigation werewere constructedconstructed fromfrom aa 1515× × 5 ×× 33 mm thick Manganin alloy element which was electronelectron beambeam weldedwelded toto twotwo 22.522.5× × 15 × 33 mm mm thick thick low low resistance resistance copper copper terminationsterminations.. This This produc produceses a a 100 100 μµΩΩ ± 5%,5%, 3 3 W W rated rated shunt shunt resistor resistor as shown in Figure1 1.. The The resistance values values of of a a batch batch of of 50 50 shunts shunts were were measured measured to tofind ﬁnd a sam a sampleple of 20 of parts 20 parts in the in range the range 95– 9995–99 μΩ µwhichΩ which were were suitable suitable for forabrasive abrasive trimming trimming to a to target a target value value of 100 of 100μΩµ. Ω.

Figure 1. ManganinManganin Alloy Alloy Shunt Shunt Construction Construction (dimensions in mm).

2.2. Abrasive Trimming System The experimental setup for the concurrent trimming process used in this study is shown in Figure 2. The shunt samples were trimmed using a Buehler Isomet 5000 linear precision saw fitted with a 178 mm diameter, 0.8 mm thick, rubber bonded silicon carbide (R/SiC) AcuThin cutting disc.

Materials 2017, 10, 876 3 of 10

2.2. Abrasive Trimming System MaterialsMaterials 20172017,, 1010,, 887676 33 ofof 1010 The experimental setup for the concurrent trimming process used in this study is shown in Figure2. TheThe shunt samplessamples samples werewere weremountedmounted trimmed inin aa PTFEPTFE using insulatedinsulated a Buehler trimming Isomettrimming 5000 jigjig linear whichwhich precision waswas securedsecured saw fitted inin thethe with machinemachine a 178 vice mmvice diameter,andand thethe cuttingcutting0.8 mm discdiscthick, waswas rubber rotatedrotated bonded atat aa speedspeed silicon ofof carbide 30003000 rpmrpm (R/SiC) andand AcuThin fedfed intointo the cuttingthe sideside disc. ofof thethe The shuntshunt samples atat aa were feedfeed mountedraterate ofof 1.21.2 in mm/minmm/min a PTFE. insulated. TheThe feedfeed trimming waswas managedmanaged jig which viavia aa was leadscrewleadscrew secured andand in the drivedrive machine motormotor vice whichwhich and waswas the cuttingdrivendriven by discby aa wasvariablevariable rotated powerpower at a speedsupplysupply of unitunit 3000,, andand rpm controlledcontrolled and fed into usingusing the sideaa programprogram of the shunt developeddeveloped at a feed inin rateLabVIEWLabVIEW of 1.2 mm/min.softwaresoftware onon The aa feedPC.PC. was managed via a leadscrew and drive motor which was driven by a variable power supply unit, and controlled using a program developed in LabVIEW software on a PC.

FigureFigure 2.2. SchemSchematicSchematicatic ofof thethe concurrentconcurrent trimmingtrimming system.system.

TheThe resistance resistance value value of of the the the shunt shunt shunt was was was continuously continuously continuously measured measured during during trimming trimming using using the the combinationcombination of of an an an Agilent Agilent Agilent B2900A B2900A B2900A source source source meter meter meter and and and 34420A 34420A 34420A nanovolt nanovolt nanovolt meter. meter. meter. All All resistance resistance resistance measurementsmeasurements were were were performed performed performed using using using the the the four-wire fourfour--wirewire Kelvin Kelvin Kelvin method; method; method; a ﬁxed a a current fixed fixed current current of 1 A was of of 1supplied 1 A A was was tosuppliedsupplied the current toto thethe (I) currentcurrent terminals (I)(I) terminalsterminals of the shunt ofof thethe by shuntshunt the B2900A byby thethe B2900AB2900A source meter, sourcesource whilst meter,meter, the whilstwhilst voltage thethe voltagevoltage drop across dropdrop theacrossacross pre-soldered thethe prepre--solderedsoldered voltage voltagevoltage (V) terminals (V)(V) terminalsterminals was continuously waswas continuouslycontinuously monitored monitoredmonitored using the usingusing 34420A thethe 34420A34420A nanovolt nanovoltnanovolt meter. Thesemeter.meter. current TheseThese currentcurrent and voltage andand voltage measurementsvoltage measurementsmeasurements were simultaneously werewere simultaneouslysimultaneously read by the readread PC byby and thethe LabVIEW PCPC andand LabVIEWLabVIEW software usingsoftwaresoftware a Keysight usingusing aa KeysightKeysight USB/GPIB USB/GPIBUSB/GPIB interface interfaceinterface and were andand used werewere to usedused calculate toto calculatecalculate the resistance thethe resistanceresistance value of valuevalue the shuntofof thethe beingshuntshunt trimmed.beingbeing trimmed.trimmed. Oncethis OnceOnce calculated thisthis calculatedcalculated resistance resistanceresistance value reached valuevalue reachedreached the target ththee value targettarget of valuevalue 100 µ Ωofof, 100100 the PCμμΩΩ, sent, thethe aPCPC signal sentsent toaa signalsignal the variable toto thethe power variablevariable supply powerpower to supplysupply reverse toto the reversereverse voltage thethe supply voltagevoltage to the supplysupply leadscrew toto thethe motor leadscrewleadscrew and retractmotormotor theandand cutting retractretract disc thethe cutting fromcutting the discdisc shunt. fromfrom Figure thethe shunt.shunt.3 shows FigureFigure side and 33 showsshows front side viewsside andand of thefrontfront shunt viewsviews during ofof thethe the shuntshunt concurrent duringduring trimmingthethe concurrentconcurrent process. ttrimmingrimming process.process.

FigureFigure 3.3. (a)((a)a) Side Side and and (b) ((b)b) FrontFront viewsviews ofof thethe concurrentconcurrent trimmingtrimming system.system.

2.32.3.. CharacterizationCharacterization FollowingFollowing trimming,trimming, thethe resistanceresistance accuracyaccuracy andand distributiondistribution ofof thethe samplessamples wwereere determineddetermined beforebefore the the shunts shunts were were subjected subjected to to a a series series of of further further tests tests to to establish establish if if trimming trimming causes causes any any

Materials 2017, 10, 876 4 of 10

2.3. Characterization Following trimming, the resistance accuracy and distribution of the samples were determined before the shunts were subjected to a series of further tests to establish if trimming causes any signiﬁcant change to the electrical properties of the components when compared to untrimmed parts. All resistance measurements were achieved using the four-wire Kelvin method and the Agilent B2900A and 34420A meters described earlier. TCR measurements were conducted in order to simulate the temperature that the shunt resistors may be subjected to whilst in a smart energy meter. The tests were performed in accordance with MIL STD 202 Method 304, using a Grant LTC1 model GD120-R2 refrigerated bath and re-circulator. Each shunt was submerged in Shell transformer oil type Diala S3 ZX-1G while its resistance value was continuously monitored. The temperature of the oil bath was then increased from −10 to +110 ◦C in 10 ◦C increments and allowed to stabilize for 10 min at each measurement point before the resistance value of the shunt was recorded. Equation (1) was then used to express the rate of change in resistance (∆R) value per 1 ◦C(TCR) in the prescribed temperature range (∆T)

∆R TCR = 10−6 (1) R × ∆T

Next, the shunt resistors were connected to a high current power supply (Glassman LPC 6-220) to determine if trimming the Manganin resistive element has any effect on the temperature rise and thus potential power rating of the shunt when used in the smart energy meter. Prior to testing, a type K thermocouple was resistance spot welded to the rear face of the Manganin element of each shunt to measure its temperature rise. The shunts were then powered up to 4 W (~200 A) at a rate of 10 A per second and allowed to stabilize until the temperature of the Manganin element reached a steady state. Thermal images were then taken using a FLIR T620bx camera. Due to the highly reflective properties of the shunt resistors materials, the samples were all painted with Pyromark 1200 high temperature black coating (ε = 0.97) prior to thermal imaging. High temperature resistance stability tests were conducted in accordance with BS EN 60115-1 standard. The initial resistance value of the shunts were measured before they were stored in an oven in ambient air for 168 h at an elevated test temperature of 200 ◦C. During testing, the resistance value of the shunts was measured at time intervals of 24, 48, 72, and 168 h and then compared with the initial resistance values to monitor any change in resistance. The morphology and composition of the Manganin shunt samples were investigated using a FEI Quanta 200 scanning electron microscope (SEM) at an accelerating voltage of 20 kV. Chemical composition was acquired using an Oxford Instruments INCA X-ray detector at 8000× magnification for 60 s. Energy dispersive spectroscopy (EDS) analysis measurements were made using a lithium-drifted silicon detector attached to the SEM. Surface roughness measurements were performed in accordance with ISO 4287 standard using an Alicona Infinite Focus microscope at ×10 magnification.

3. Results and Discussion Top and side views of a typical square cut trim geometry of a Manganin shunt resistor element are shown in Figure4a,b respectively. The depth of cut for all samples was in the range of 0.5 to 1.2 mm depending on the initial resistance value of the part which equates to around 3 to 8% of the initial Manganin element width of 15 mm. The edges of the trimmed grooves were found to be relatively free from burrs and the surface of the grooves were smooth, having roughness average (Ra) values in the range 100.85 to 133.95 nm. Figure5 shows the measured resistance value of the 20 samples before (untrimmed) and after trimming with a square cut to a target resistance value of 100 µΩ. It can be clearly seen that the resistance distribution of the sample of shunts after trimming is much smaller than that in the Materials 2017, 10, 876 5 of 10 untrimmed condition with values of 95.28 to 98.82 µΩ and 99.96 to 100.56 µΩ, respectively. The resistance values of the trimmed parts are also all well within ±1% of the target resistance value of 100 µMaterialsΩ, thus 2017 highlighting,, 10,, 876 the system’s capability to accurately trim the shunt resistors. 5 of 10

Materials 2017, 10, 876 5 of 10

FigureFigure 4. (4.a) (a Top) Top and and (b ()b side) side views views ofof aa shunt resistor resistor trimmed trimmed with with a square a square cut. cut. Figure 4. (a) Top and (b) side views of a shunt resistor trimmed with a square cut.

Figure 5. Resistance values of the shunt resistors before (untrimmed) and after trimming with the FigureFigure 5. Resistance 5. Resistance values values of of the the shunt shunt resistorsresistors before (untrimmed) (untrimmed) and and after after trimming trimming with with the the square cut. squareFigure cut. 5. Resistance values of the shunt resistors before (untrimmed) and after trimming with the square cut. 3.1.. Temperature Co-Efficient of Resistance (TCR) 3.1. Temperature Co-Efﬁcient of Resistance (TCR) 3.1. TemperatureThe graph ofCo -resistanceEfficient of againstResistance temperature (TCR) in Figure 6 shows a positive correlation for both untrimmedThe graph and of resistance square cut againstshunt resistors temperature in the temperature in Figure6 rangeshows of a−10 positive to +110 correlation °C. This positive for both The graph of resistance against temperature in Figure 6 shows a positive correlation◦ for both untrimmedTCR, or decrease and square in conductivity cut shunt resistors with an increase in the temperature in temperature, range is typical of −10 in tometals +110 andC. alloys This and positive untrimmed and square cut shunt resistors in the temperature range of −10 to +110 °C. This positive TCR,can or be decrease related in to conductivityan increase in with energy an with increase temperature in temperature, which causes is typical the ions in to metals vibrate and more alloys TCR, or decrease in conductivity with an increase in temperature, is typical in metals and alloys and frequently and subsequently impede the conduction of electrons [10].. and can be related to to an an increa increasese in in energy energy with with temperature temperature which which causes causes the the ions ions to vibrate to vibrate more more frequentlyfrequently and and subsequently subsequently impede impede the the conduction conduction of electrons electrons [10] [10. ].

Figure 6. Resistance versus temperature for untrimmed and square cut trimmed shunt resistor samples. Figure 6. Resistance versus temperature for untrimmed and square cut trimmed shunt resistor Figure 6. Resistance versus temperature for untrimmed and square cut trimmed shunt resistor samples. samples.

Materials 2017, 10, 876 6 of 10

It can be further observed that the magnitude of this positive TCR is relatively small across this Materials 2017, 10, 876 6 of 10 broadMaterials temperature 2017, 10, 876 range of −10 to +110 ◦C, which is a propitious property in the manufacture6 of 10 of precisionI shuntt can be resistors. further observed The calculated that the TCRmagnitude values of were this positive consistent TCR across is relatively the full small temperature across this range It can be further observed that the magnitude of this positive TCR is relatively small across this for bothbroad the temperature samples ofrange five of untrimmed −10 to +110 and°C, which five square is a propitious cut trimmed property parts, in the having manufacture values of in the broad temperature range of −10 to +110 °C, which is a propitious property in the manufacture of rangeprecision of +105 shunt to +108 resistors ppm/. The◦C calculated and +92 to TCR +94 values ppm/ were◦C, respectively.consistent across The the average full temperature TCR value range for the precision shunt resistors. The calculated TCR values were consistent across the full temperature range for both the samples of five untrimmed◦ and five square cut trimmed parts, having values in the range ◦ untrimmedfor both resistorsthe samples was of 106 five ppm/ untrimmedC whereas and five that square for thecut abrasivetrimmed parts, trimmed having resistors values was in the 93 range ppm/ C. of +105 to +108 ppm/°C and +92 to +94 ppm/°C, respectively. The average TCR value for the◦ Theseof TCR +105 values to +108 are ppm/°C higher and than +92 those to +94 previously ppm/°C, reportedrespectively. for The pure average Manganin TCR of value 15 ppm/ for theC[ 4], untrimmed resistors was 106 ppm/°C whereas that for the abrasive trimmed resistors was 93 ppm/°C. ◦ withuntrimmed this increase resistors being was related 106 ppm/°C to the whereas series resistancethat for the abrasive and highly trimmed positive resistors TCR was (+3930 93 ppm/°C. ppm/ C) These TCR values are higher than those previously reported for pure Manganin of 15 ppm/°C [4], contributionThese TCR of values the two are small higher areas than ofthose copper previously termination reported on for either pure sideManganin of the of Manganin 15 ppm/°C resistive [4], with this increase being related to the series resistance and highly positive TCR (+3930 ppm/°C) with this increase being related to the series resistance and highly positive TCR (+3930 ppm/°C) elementcontribution [11]. However, of the two the small difference areas inof TCRcopper between termination the untrimmed on either side and of trimmed the Manganin parts, resistive which is the contribution of the two small areas of copper termination on either◦ side of the Manganin resistive mainelement purpose [11] of. thisHowever, study, the is relativelydifference small,in TCR at between around the 13.4 untrimmed ppm/ C and across trimmed the full parts, test which temperature is element [11]. However,◦ the difference in TCR between the untrimmed and trimmed parts, which is rangethe of main−10 to purpose +110 ofC, this thus study suggesting, is relatively that trimming small, at has around a negligible 13.4 ppm/°C effect across and in the fact full results test in the main purpose of this study, is relatively small, at around 13.4 ppm/°C across the full test a slighttemperature improvement range inof −10 this to key +110 property °C, thus ofsuggesting the shunts. that trimming has a negligible effect and in fact temperature range of −10 to +110 °C, thus suggesting that trimming has a negligible effect and in fact results in a slight improvement in this key property of the shunts. 3.2. Highresults Power in a slight Performance improvement in this key property of the shunts. 3.2. High Power Performance 3.2Plots. High of Power typical Performance temperature increase in untrimmed and square cut trimmed shunt resistors, poweredP uplots to of 4 t Wypical (~200 temperature A), are shown increase in Figures in untrimmed7 and8 respectively. and square cut The trimmed results shunt for both resistors, types of Plots of typical temperature increase in untrimmed and square cut trimmed shunt resistors, powered up to 4 W (~200 A), are shown in Figures 7 and 8 respectively. The◦ results for both types of shuntpowered follow up a very to 4 W similar (~200 A), trend. are shown Starting in Fig ature rooms 7 and temperature 8 respectively. (~23 TheC) results the temperatures for both types of of both shunt follow a very similar trend. Starting at room temperature (~23 °C) the temperatures of both the the untrimmedshunt follow anda very trimmed similar trend resistor. Starting elements at room increase temperature rapidly (~23 in °C) the the first temperatures 150 s to valuesof both aroundthe ◦untrimmed ◦and trimmed resistor elements increase rapidly in the first 150 s to values around 148 °C 148 untrimmedC and 142 andC respectively. trimmed resistor After elements which, increase the increases rapidly are in more the first gradual 150 s to before values reaching around 148 maximum °C and 142 °C respectively. After which, the increases are more gradual before reaching maximum temperaturesand 142 °C of respectively. 194 ◦C and After203 ◦ C which, respectively the increases at around are more 500 s.gradual The temperature before reaching of the maximum two shunts temperatures of 194 °C and 203 °C respectively at around 500 s. The temperature of the two shunts temperatures of 194 °C and 203 °C respectively◦ at around 500 s. The temperature◦ of the two shunts thenthen decreases decreases slightly slightly to to around around 192 192 C°C forfor thethe untrimmeduntrimmed and and 198 198 °C Cfor for the the trimmed trimmed parts parts and and then decreases slightly to around 192 °C for the untrimmed and 198 °C for the trimmed parts and fluctuatefluctuate around around this this point point until until the the power power supplysupply is switched at at a atime time of of1200 1200 s. After s. After which, which, bothboth fluctuate around this point until the power supply is switched at a time of 1200 s. After which, both shuntsshunts gradually gradually cool cool back back down down to to room room temperature temperature at around around 1500 1500 s (not s (not shown shown on onplots). plots). shunts gradually cool back down to room temperature at around 1500 s (not shown on plots).

Figure 7. Temperature and resistance versus time plots for the untrimmed shunt resistor. FigureFigure 7. Temperature 7. Temperature and and resistance resistance versusversus time plots plots for for the the untrimmed untrimmed shunt shunt resistor. resistor.

Figure 8. Temperature and resistance versus time plots for the shunt resistor trimmed with the square Figure 8. Temperature and resistance versus time plots for the shunt resistor trimmed with the square Figurecut. 8. Temperature and resistance versus time plots for the shunt resistor trimmed with the cut. square cut.

Materials 2017, 10, 876 7 of 10 Materials 2017, 10, 876 7 of 10

Plots of resistance versus time for the untrimmed and trimmed shunts, powered up to 4 W, are also shown in FiguresFigures7 7 and and8 8and and follow follow a a very very similar similar trend trend to to those those described described for for the the temperature temperature rise test. The re resistancesistance value of the untrimmed part starts at 96.9 μµΩΩ and rises quickly to 97.8 µμΩΩ within approximately 200 s of thethe currentcurrent beingbeing supplied.supplied. The resistance value of the trimmed part starts at 100.3 µμΩΩ and rises in an almost identical fashion to a value of 101.4 µμΩΩ after 200200 s.s. After this point,point, the resistance values of both shunts increase more gradually before reaching maximum values of 97.9 µμΩΩ for the untrimmed part and 101.6 µμΩΩ for the trimmed part at around 500 s s.. After After this this point point,, the resistance valuesvalues ofof both shunts reduce very slightly to around 97.8 µμΩΩ and 101.5 μµΩΩ respectively and remain there until the power is removed at a time of 1200 s.s. Finally, the resistance of both shunts was observed to gradually return toto their originaloriginal values as they cool back down to room temperature at around 1500 s (not shownshown onon plots).plots). There are several important observations that can be drawn from from the the results results of of Fig Figuresures 77 andand8 8.. Firstly, thethe increaseincrease inin resistanceresistance withwith increaseincrease inin suppliedsupplied current isis signiﬁcantsignificant forfor bothboth untrimmed and trimmed parts, parts, having having a a maximum maximum value value of of around around 1 to 1 to1.3 1.3 μΩµ Ω(+1(+1 to 1.3%). to 1.3%). These These increases increases can canbe quite be quite clearly clearly related related to theto the inherent inherent TCR TCR of of the the Manganin Manganin material material discussed discussed in in Section Section 3.1;3.1; multiplying the measuredmeasured TCRTCR ofof approximatelyapproximately +100+100 ppm/ppm/°C◦C by by the the maximum temperature rise of 180 ◦°CC givesgives anan increaseincrease inin resistance resistance of of 1.8%. 1.8%. Although Although this this change change can can be be deemed deemed non-permanent, non-permanent, as theas the resistance resistance of the of shunts the shunts return return to their to original their original values once values the once power the is removed, power is this removed, temporary this increasetemporary in resistanceincrease in could resistance in turn could lead toin errorsturn lead when to measuringerrors when current measuring ﬂow incurrent a smart flow energy in a meter.smart Theenergy second meter. observation The second to observation be made is to that be trimmingmade is that the trimming shunt appears the shunt to give appears a slight to give increase a slight in theincrease peak in temperature the peak temperature rise of the resistor rise of the element. resistor This element. is further This supported is further supported by the thermal by the images thermal of theimages untrimmed of the untrimmed and trimmed and parts trimmed in Figure parts9, taken in Fig afterure 9, 500 taken s when after the 500 shunts s when had the reached shunts their had maximumreached their temperature. maximum temperature.

Figure 9. Thermal images of the shunt resistors (a) untrimmed and (b) trimmed with the square cut. Figure 9. Thermal images of the shunt resistors (a) untrimmed and (b) trimmed with the square cut. The temperature was measured at two spots on either side of the centre of the resistor element for bothThe thetemperature untrimmed was and measured square cut at trimmedtwo spots parts. on either For theside untrimmed of the centre part of inthe Figure resistor9a, element at spot 1,for the both temperature the untrimmed was 189 and◦ Csquare and at cut spot trimmed 2, it was parts. 194 ◦ForC. Onthe average,untrimmed the part temperature in Figure was 9a, 192at spot◦C. In1, the contrast temperature to the untrimmed was 189 °C part,and at for spot the 2, square it was trimmed194 °C. On part average, in Figure the9 temperatureb, at spot 1, whichwas 192 was °C. locatedIn contrast next to to the the untrimmed end of the trim part, plunge, for the the square temperature trimmed was part 198 in◦C Fig andure at 9b, spot at 2,spot it was 1, which 204 ◦C. was On average,located next the temperatureto the end ofwas the 202trim◦ plunge,C, which the is 10temperature◦C higher thanwas 198 that °C for and the at untrimmed spot 2, it was shunt. 204 This °C. increaseOn average, in temperature the temperature for the was trimmed 202 °C, shuntwhich is is comparable 10 °C higher with than that that reported for the untrimmed in Figures7 andshunt.8 andThis mayincrease be attributed in temperature to the for square the trimmed shaped plungeshunt is cut comparable which impedes with that the reported ﬂow of currentin Figure throughs 7 and the8 and resistive may be material attributed and to causes the square local currentshaped crowdingplunge cut in which the central impedes region the offlow the of element current [ 12through]. This inferiorthe resistive performance material couldand causes be a cause local forcurrent concern crowding as increased in the operating central region temperature of the element is known [12] to. This lead toinferior accelerated performance ageing andcould premature be a cause failure for concern of resistive as increased devices [13 operating]. This issue temperature is discussed is known further into thelead following to accelerated section. ageing and premature failure of resistive devices [13]. This issue is discussed further in the following section.

3.3. High Temperature Performance Results of average change in resistance of both the samples of five untrimmed and five square cut trimmed shunt samples following storage for 168 h at 200 °C in air are shown in Figure 10 with

Materials 2017, 10, 876 8 of 10

3.3. High Temperature Performance

MaterialsResults 2017, 10 of, 8 average76 change in resistance of both the samples of five untrimmed and five square8 of cut 10 trimmed shunt samples following storage for 168 h at 200 ◦C in air are shown in Figure 10 with interim resultsinterim reported results reported at time intervals at time inte of 24,rvals 48, of and 24, 72 48, h. and The 72 graph h. The shows graph that shows the average that the resistance average valuesresistance of both values the of untrimmed both the untrimmed and square and cut square trimmed cut shunt trimmed samples shunt increase samples significantly increase significantly in the first 24in the h of first the test,24 h withof the changes test, with of 1.7%changes and of 2.3% 1.7% respectively. and 2.3% respectively. This rate of changeThis rate reduces of change significantly reduces insignificantly the following in the 24 following h of testing 24 and h of both testing samples and both reach samples a maximum reach valuea maximum of change value in resistanceof change ofin 2.5%resistance after 48of h2.5% of storage. after 48 Afterh of storage. this point, After the this resistance point, the value resistance of both shuntsvalue of decreases both shunts following decreases 72 h offollowing storage 72 before h of increasingstorage before again increasing to give final again changes to give infinal resistance changes values in resistance after 168 values h of 1.85%after 168 and h 2.29%of 1.85% for and the 2.29% untrimmed for the and untrimmed square cut and trimmed square shuntcut trimmed samples shunt respectively. samples respectively.

Figure 10. 10. AverageAverage change change in resistancein resistance for untrimmed for untrimmed and square and square cut trimmed cut trimmed shunt resistor shunt samples resistor duringsamples storage during at storage 200 ◦C at for 200 168 °C h. for 168 h.

This initial increase and then subsequent decrease in resistance can be related to two competing This initial increase and then subsequent decrease in resistance can be related to two competing mechanisms. Initial oxidation of the surface of the Manganin element, resulting in the formation of a mechanisms. Initial oxidation of the surface of the Manganin element, resulting in the formation high resistance oxide layer, which in turn increases the overall resistance of the material, and of a high resistance oxide layer, which in turn increases the overall resistance of the material, and annealing out of impurities and grain boundary reduction both of which lead to a reduction in annealing out of impurities and grain boundary reduction both of which lead to a reduction in resistance of the material [14]. The higher rate of change of resistance in the first 24 h for the trimmed resistance of the material [14]. The higher rate of change of resistance in the ﬁrst 24 h for the trimmed shunt can be related to an increased level of surface oxidation of the freshly exposed Manganin shunt can be related to an increased level of surface oxidation of the freshly exposed Manganin surrounding the trim cut. This theory is supported by the EDS chemical analysis of oxygen on the surrounding the trim cut. This theory is supported by the EDS chemical analysis of oxygen on the shunts’ surfaces, a summary of the results of which are reported in Table 1. shunts’ surfaces, a summary of the results of which are reported in Table1.

Table 1. EDS chemical analysis for oxygen content of untrimmed and square cut trimmed shunt Table 1. EDS chemical analysis for oxygen content of untrimmed and square cut trimmed shunt resistor resistor samples before and after the high temperature stability test samples before and after the high temperature stability test wt.% Oxygen Spectrum No. Untrimmedwt % OxygenSquare Cut Spectrum No. BeforeUntrimmed After Before Square CutAfter 1 Before4.42 After10.49 Before6.05 13.95 After 12 4.422.01 10.4916.72 6.053.60 18.53 13.95 23 2.012.29 16.726.81 3.603.61 13.87 18.53 34 2.293.08 6.8114.24 3.615.45 15.39 13.87 45 3.084.60 14.249.16 5.456.71 9.83 15.39 5 4.60 9.16 6.71 9.83 66 4.174.17 14.0914.09 7.767.76 16.69 16.69 Average 3.43 11.92 5.53 14.71 Average 3.43 11.92 5.53 14.71

The results show that the untrimmed and square cut trimmed samples have similar average levels of oxygen concentration prior to testing at 3.43 wt.% and 5.53 wt.% respectively. However, these levels increase significantly for both samples after storage at 200 °C for 168 h with values of 11.92 wt.% and 14.71 wt.% respectively.

Materials 2017, 10, 876 9 of 10

The results show that the untrimmed and square cut trimmed samples have similar average levels of oxygen concentration prior to testing at 3.43 wt % and 5.53 wt % respectively. However, these levels increase signiﬁcantly for both samples after storage at 200 ◦C for 168 h with values of 11.92 wt % and Materials 2017, 10, 876 9 of 10 14.71 wt % respectively. These results are further supported by the images of the untrimmed parts before and after testing shown in FigureFigure 1111,, which clearly show a discoloration and formation of a surface oxide after high temperature exposure.

(a)

(b)

Figure 11. ImagesImages of of shunt shunt resistor resistor ( a) before and (b) after the high temperature stabilitystability test.test.

This increase in surface oxidation of the Manganin element at the elevated test temperature leads This increase in surface oxidation of the Manganin element at the elevated test temperature leads to an increase in resistance value of the shunts. Although the resistance change of the square cut to an increase in resistance value of the shunts. Although the resistance change of the square cut trimmed sample is slightly higher at 2.29%, the stability performance of the untrimmed part is also trimmed sample is slightly higher at 2.29%, the stability performance of the untrimmed part is also unacceptable at 1.85%. This is an area that needs to be addressed before focusing on reducing the unacceptable at 1.85%. This is an area that needs to be addressed before focusing on reducing the additional effect of trimming on this key performance characteristic of the shunts. additional effect of trimming on this key performance characteristic of the shunts.

4.4. Conclusions Work in in this this paper pape hasr has demonstrated demonstrated that abrasivethat abrasive trimming trimming has the has ability theto ability reduce to the reduce resistance the resistance tolerance of 100 μΩ Manganin shunt resistors from typical values of ±5% to less than ±1% tolerance of 100 µΩ Manganin shunt resistors from typical values of ±5% to less than ±1% whilst whilst having minimal effect on the key electrical properties of the device. having minimal effect on the key electrical properties of the device. The TCR, high current, and high temperature performance of the square cut trimmed shunts The TCR, high current, and high temperature performance of the square cut trimmed shunts were compared to that of untrimmed shunts. It was found that the average TCR value was slightly were compared to that of untrimmed shunts. It was found that the average TCR value was slightly reduced following trimming with typical results of +106 ppm/°C and +93 ppm/°C for untrimmed and reduced following trimming with typical results of +106 ppm/◦C and +93 ppm/◦C for untrimmed trimmed shunts respectively. When subjected to a high current of 200 A the trimmed part showed a and trimmed shunts respectively. When subjected to a high current of 200 A the trimmed part showed slight increase in temperature to 203 °C as compared to 194 °C for the untrimmed part but both parts a slight increase in temperature to 203 ◦C as compared to 194 ◦C for the untrimmed part but both parts had significant temporary increases in resistance of around 1 to 1.3 μΩ. This result was quite had significant temporary increases in resistance of around 1 to 1.3 µΩ. This result was quite accurately accurately explained using the inherent TCR property of the Manganin material. The results for explained using the inherent TCR property of the Manganin material. The results for resistance change resistance change following high temperature storage at 200 °C for 168 h were also significant for following high temperature storage at 200 ◦C for 168 h were also significant for both untrimmed and both untrimmed and trimmed parts with shifts of 1.85 and 2.29% respectively and these results were trimmed parts with shifts of 1.85 and 2.29% respectively and these results were related to surface related to surface oxidation of the shunts which was accelerated for the freshly exposed surfaces of oxidation of the shunts which was accelerated for the freshly exposed surfaces of the trimmed part. the trimmed part. Although this work has demonstrated that trimming does not have any detrimental effect on Although this work has demonstrated that trimming does not have any detrimental effect on the electrical performance of the Manganin shunt resistors it has in turn highlighted some significant the electrical performance of the Manganin shunt resistors it has in turn highlighted some significant inherent performance issues with the Manganin resistive alloy material itself. Therefore, future work inherent performance issues with the Manganin resistive alloy material itself. Therefore, future work should focus on improving the high temperature performance of this material as well as further should focus on improving the high temperature performance of this material as well as further developing the abrasive trimming process itself to produce accurate and reliable shunt resistors for developing the abrasive trimming process itself to produce accurate and reliable shunt resistors for use in high power smart metering applications. use in high power smart metering applications.

Acknowledgments: This work was funded and supported by Majlis Amanah Rakyat (MARA), Malaysia.

Author Contributions: S.N.M. performed the experiments and wrote the draft manuscript. M.B. analysed the data for TCR and edited the manuscript. R.P. analysed the data for high power performance. D.B. collaborated in the editing process.

Conflicts of Interest: The authors declare no conflict of interest.

Materials 2017, 10, 876 10 of 10

Acknowledgments: This work was funded and supported by Majlis Amanah Rakyat (MARA), Malaysia. Author Contributions: S.N.M. performed the experiments and wrote the draft manuscript. M.B. analysed the data for TCR and edited the manuscript. R.P. analysed the data for high power performance. D.B. collaborated in the editing process. Conﬂicts of Interest: The authors declare no conﬂict of interest.

References

1. O’Driscoll, E.; O’Donnell, G.E. Industrial power and energy metering—A state-of-the-art review. J. Clean. Prod. 2013, 41, 53–64. [CrossRef] 2. Manov, V.; Adar, E.; Geller, M.; Sorkine, E.; Margolin, I. Amorphous Metallic Alloy Electrical Heater Systems. U.S. Patent 5,641,421, 24 June 1997. 3. Haynes, W.M. CRC Handbook of Chemistry and Physics; CRC Press: Boca Raton, FL, USA, 2014. 4. Casey, K. The Need for Current Sense, Circuit Protection and Communications in Smart Meters; Electronic Design: New York, NY, USA, 2011. 5. Weranga, K.S.K.; Chandima, D.P.; Kumarawadu, S.P. Smart metering for next generation energy efﬁciency & conservation. In Proceedings of the 2012 IEEE Innovative Smart Grid Technologies—Asia (ISGT Asia), Tianjin, China, 21–24 May 2012. 6. Isabellenhütte Company. Low-Ohmic Precision and Power Resistors; Isabellenhütte Company: Dillenburg, Germany, 2012; p. 14. 7. Alafogianni, M.; Penlington, R.; Birkett, M. Resistor trimming geometry; past, present and future. In IOP Conference Series: Materials Science and Engineering; IOP Publishing: Bristol, UK, 2016. 8. Sandborn, P.; Sandborn, P.A. A Random Trimming Approach for Obtaining High-Precision Embedded Resistors. IEEE Trans. Adv. Packag. 2008, 31, 76–81. [CrossRef] 9. Misti, S.N.; Birkett, M.; Bell, D.; Penlington, R. Practical Implementation of a Concurrent Trimming Approach for Manganin Shunt Resistors. Int. J. Comput. Commun. Instrum. Eng. 2015, 2, 81–85. 10. Meaden, G.T. Electrical Resistance of Metals; Springer: New York, NY, USA, 2013. 11. Dyos, G. The Handbook of Electrical Resistivity: New Materials and Pressure Effects; IET Materials and Device Series 13; The Institution of Engineering and Technology: London, UK, 2012. 12. Zhu, X.; Kotadia, H.; Xu, S.; Lu, H.; Mannan, S.; Bailey, C.; Chan, Y. Modeling Electromigration for Microelectronics Design. J. Comput. Sci. Technol. 2013, 7, 251–264. [CrossRef] 13. Ohring, M. Reliability and Failure of Electronic Materials and Devices; Academic Press: Cambridge, MA, USA, 1998. 14. Reed-Hill, R.E.; Abbaschian, R. Physical Metallurgy Principles; Van Nostrand: New York, NY, USA, 1973.

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).